Does a lighter deload week cost gains?
In 19 untrained men, two 2-set weeks in an 8-week lifting block showed no evidence of a difference from training straight through for muscle and 10RM gains. Endurance untested.
Pancar and colleagues asked whether a deload made of fewer sets and fewer sessions, rather than a week off, costs anything in an 8-week resistance programme.
Nineteen untrained young men trained unilateral leg extension and unilateral dumbbell biceps curl to technical failure at 8-12RM; for each participant one leg and one arm were randomly assigned to continuous training (twice a week, 6-8 sets per week per exercise) and the other leg and arm to the same programme with deload weeks in weeks 4 and 8 (one session, 2 sets per exercise).
Over the 8 weeks the deload limbs did 46 sets per muscle group against 56. Both conditions increased muscle thickness at all seven ultrasound sites and 10RM in both exercises, with one exception (lateral quadriceps at 30% in the deload limb, p = 0.073); no time × condition interaction reached significance (p = 0.239-0.955) and every 95% CI of the difference included zero.
The authors conclude that a reduced-volume, reduced-frequency deload at the midpoint and end of a short programme does not appear to hinder hypertrophy or strength-endurance in untrained men.
Our reading: no evidence of a cost from two lighter weeks, measured in a population that gains from very little volume; every muscle-thickness point estimate favoured continuous training (by up to about 1 mm, 16-20% of the gain at the mid-thigh sites, our arithmetic), the CIs are wide enough to hide a moderate cost, and the week-8 deload doubled as a light week before testing.
It contrasts with REF-217, where a full week off in trained lifters cost some lower-body strength, but the two trials differ in population, deload type and outcome.
Protocol
Design: randomized within-subject trial; legs and arms randomly assigned to the two conditions by a computer-generated permuted block sequence (block size 4), concealed until assignment, generated by an independent researcher and applied by a researcher blinded to allocation; outcome assessors and statisticians blinded, participants not. 11 weeks: 1 familiarization, 1 pre-test, 8 training, 1 post-test.Ethics: Gaziantep University (2024/03). Registered on OSF (12 January 2025) and ClinicalTrials.gov (NCT06825052, 13 February 2025); the data-availability statement mentions a "Stage 2 manuscript", the wording of a registered report, but the article does not call itself one.
Participants: 20 recruited, 1 dropped out with an unrelated injury, n = 19 analysed; men aged 18-35 with no structured resistance training in the previous 6 months, free of disorders, PAR-Q negative, no anabolic agents; age 21.79 ± 1.69 y, height 170.84 ± 8.95 cm, body mass 63.21 ± 9.99 kg, BMI 21.62 ± 2.46. Habitual diet kept; creatine not allowed.
Continuous condition: each exercise twice a week, 6-8 sets per week per exercise, 8-12 repetitions to technical failure; load changed by about 10% from set to set to stay in range; tempo about 1 s concentric and 2 s eccentric (metronome); 120 s between sets, 30 s between limbs; the limb that started alternated by week.
Deload condition: the same programme except weeks 4 and 8: one session a week with 2 sets per exercise (sets per exercise cut by about 66-75%, frequency from two to one). Total over 8 weeks: 46 vs 56 sets per muscle group (about 18% fewer).
Outcomes: muscle thickness by B-mode ultrasound (Siemens Acuson S2000), elbow flexors at 60% of the epicondyle-acromion distance, mid-thigh and lateral thigh at 30%, 50% and 70% of femur length; three images per site averaged; the same technician, blinded, at least 48 h after training. Strength-endurance: 10RM for each exercise (NSCA procedure), repeated 48-72 h later, the higher load taken.
Analysis: 2 × 2 repeated-measures ANOVA (time × condition), paired change-score differences with 95% CIs, Bonferroni post hoc tests, partial eta squared (0.01 small, 0.06 medium, 0.14 large); complementary one-sided Bayesian paired t-tests (JZS prior) within each condition; Jamovi 2.5.6.0.
Sample size: a priori for power 0.8 and alpha 0.05 with d = 0.55 taken from hypertrophy meta-analyses, minimum 15; the authors note that the observed pre-post effects exceeded 0.55, so the calculation addresses training effects, not the between-condition difference.
Funding and interests: Princess Nourah bint Abdulrahman University Researchers Supporting Project (PNURSP2026R424), funders without a role in design, analysis or publication; the authors declare no competing interests.
Key Results
Change from pre to post in each condition and the paired difference (Table 1; mm for muscle thickness, kg for 10RM):
| Outcome | Change, deload (95% CI) | Change, continuous (95% CI) | Difference, deload minus continuous (95% CI) | p (paired difference) | Interaction effect size (partial eta squared) |
|---|---|---|---|---|---|
| Lateral quadriceps 30% | +1.99 (−0.76 to 4.74) | +2.27 (0.06 to 4.48) | −0.28 (−2.61 to 2.05) | 0.803 | 0.002 |
| Lateral quadriceps 50% | +3.18 (1.91 to 4.45) | +3.94 (2.60 to 5.28) | −0.76 (−2.20 to 0.68) | 0.282 | 0.030 |
| Lateral quadriceps 70% | +4.12 (2.11 to 6.13) | +4.20 (2.48 to 5.92) | −0.08 (−2.12 to 1.96) | 0.937 | 0.001 |
| Mid quadriceps 30% | +4.16 (2.97 to 5.36) | +5.09 (3.71 to 6.47) | −0.93 (−2.34 to 0.48) | 0.183 | 0.058 |
| Mid quadriceps 50% | +5.67 (4.18 to 7.15) | +6.74 (5.06 to 8.43) | −1.08 (−2.75 to 0.59) | 0.191 | 0.057 |
| Mid quadriceps 70% | +4.46 (3.38 to 5.54) | +5.61 (4.11 to 7.11) | −1.15 (−2.63 to 0.33) | 0.119 | 0.076 |
| Upper arm 60% | +6.03 (4.48 to 7.58) | +6.33 (4.69 to 7.96) | −0.30 (−1.88 to 1.29) | 0.698 | 0.004 |
| 10RM leg extension | +17.63 (12.66 to 22.61) | +16.84 (11.75 to 21.94) | +0.79 (−0.66 to 2.24) | 0.268 | 0.068 |
| 10RM biceps curl | +3.55 (2.47 to 4.64) | +3.42 (2.42 to 4.42) | +0.13 (−0.14 to 0.41) | 0.331 | 0.004 |
Within each condition: every site and both 10RM tests rose significantly in both conditions except the lateral quadriceps at 30% in the deload limb (t = −1.52, p = 0.073, d = −0.349 with a CI including zero, Bayes factor 1.163, "anecdotal"); the other within-condition Bayes factors were mostly above 1000.
Direction: all seven muscle-thickness differences were negative, the continuous limb ahead by 0.08 to 1.15 mm; at the three mid-thigh sites the deload limb gained 16-20% less than the continuous limb (our arithmetic), none significant. Both 10RM differences favoured the deload limb by under 1 kg.
Precision: the CIs allow the continuous limb to have gained up to 2.63 mm more at mid-thigh 70%, about half of its own gain (our arithmetic). No equivalence test was run, so the honest reading is no evidence of a difference, not equivalence.
Reporting: the abstract calls the between-condition effects small (partial eta squared 0.001-0.076) and the Results call 0.076 "negligible in magnitude", while the paper's own thresholds put 0.06 at medium: two outcomes (10RM leg 0.068, mid-thigh 70% 0.076) are medium by those thresholds and two more (0.057, 0.058) sit just below.The abstract's "6–8 sets per exercise" is per week in the Methods. Table 2 labels the arm site "70% anterior upper arm" while the Methods and Table 1 give 60%.
Mechanisms
Measured: muscle thickness and 10RM rose in both conditions with no time × condition interaction on any outcome (p = 0.239-0.955).
Authors' interpretation: cutting volume and frequency rather than stopping may protect against the losses reported after short breaks (they cite the full-week-off trial, REF-217, and short detraining studies) while saving time, 18% fewer sets for similar gains; the 10RM test resembled the 8-12RM training, which may have eased transfer from training to test.
Inference: the week-8 deload was the final training week before post-testing, so in that half of the design the deload limb was tested after a lighter week, like a short taper. A light final week could hide a small cost of the week-4 deload or simply freshen that limb for testing; the design cannot separate the two.
Inference: untrained lifters gain from very little volume (the authors cite maintenance at one-third and one-ninth of the training volume), so removing a few sets in two of eight weeks may cost little in anyone this untrained; whether it holds for trained athletes near their ceiling was not tested.
Inference: in a within-subject design cross-education (the authors cite about 7%) and whole-body effects pull the two limbs together, which biases toward no difference, most for strength.
Practical Application
Does a lighter deload week cost gains? In 19 untrained men, two 2-set weeks in an 8-week lifting block showed no evidence of a difference from training straight through for muscle and 10RM gains. Endurance untested.
Lighter rather than off (inference): in trained lifters a full week off cost some lower-body strength (REF-217, B); here a reduced week in untrained lifters cost nothing measurable. The trials differ in population, deload type and outcome and nobody compared the two deload types head-on, so "if you deload, reduce rather than stop" is a reading of two trials, not a tested rule.
For an endurance recovery week (transfer): there is no endurance outcome here. It makes a planned reduced-volume week unlikely to cost much over a short block; it gives no evidence that the week helps either, since neither condition gained more. When to take one (signals first, or a planned lighter week) stays house practice (SYNTH-29).
Scheduled, not triggered: both deloads sat at fixed weeks; a signal-triggered deload was not tested, so this neither supports nor undercuts the base's signal rule.
Time-pressed athletes in a strength block: 18% fewer sets gave similar gains in untrained men (B+, 8 weeks); one lighter week in a busy month is a cheap trade (inference).
Limitations
Population and task: untrained young men, two single-joint unilateral exercises, 8 weeks; the authors say the result cannot be generalized to trained people, adolescents or older adults, nor to programmes with more exercises or higher volume.
Bearing on [SYNTH-29]: whether an endurance deload (reduced volume, not a week off) carries any cost is still untested; this trial answers it only for resistance training in untrained men.
Power: the sample size was computed for pre-post training effects (d = 0.55, minimum 15), not for a difference between conditions; differences of about 1 mm at mid-thigh were not significant and could be real.
Within-subject design: cross-education and systemic effects can narrow a difference between limbs, most for strength (the authors' limitation).
No time-matched non-exercise control and a single pre-test ultrasound session, so measurement drift cannot be quantified (the authors' limitation).
Timing: the week-8 deload preceded the post-tests, acting partly as a light week before testing.
Strength measure: 10RM is strength-endurance close to the training range, not maximal strength.
Reporting: "small" in the abstract for effect sizes that reach the paper's own medium threshold; the abstract's "6–8 sets per exercise" is per week.
Funding: a university researchers' support grant; no competing interests declared.
How this entry was checked and changed
Why this level
randomized within-subject design, n = 19, read from the full text (open access, CC BY-NC-ND 4.0). Each participant is their own control, so the rubric's crossover band (15 or more) applies; within B+ a weak one: untrained men, two single-joint exercises, 8 weeks, no non-exercise control, cross-education between limbs can narrow a difference, and the sample size was computed for training effects rather than for a difference between the conditions
Full text (checked 2026-10-27)
The final sample size was 19 participants (age = 21.79 ± 1.69; height = 170.84 ± 8.95 cm; weight = 63.21 ± 9.99) (full text)
Over the 8-week intervention period, the deload protocol involved approximately 18% lower total set volume per muscle group compared to the continuous protocol (deload: 46 sets, continuous: 56 sets) (full text)
Mid quadriceps 70%: deload + 4.46 mm [3.38, 5.54], continuous + 5.61 mm [4.11, 7.11], paired difference − 1.15 mm [− 2.63, 0.33], p = 0.119; interaction ηp2 = 0.076 (full text)
10RM leg extension: deload + 17.63 kg [12.66, 22.61], continuous + 16.84 kg [11.75, 21.94], paired difference + 0.79 kg [− 0.66, 2.24], p = 0.268 (full text)
Effect sizes were interpreted with thresholds of 0.01 (small), 0.06 (medium), and 0.14 (large); the abstract calls the between-condition effects small (ηp2 = 0.001–0.076) (full text)
Sample size: a statistical power of 0.8 and an α level of 0.05, with an estimated minimum of 15 participants, using Cohen's d = 0.55 (full text)
The cross-education effect is typically around 7% (full text)
Funded by the Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2026R424); the authors declare no competing interests (full text)
Content correction 2026-10-27
Correction 2026-10-27: pointers to SYNTH-29 aligned with its 2026-10-27 rewrite (single-run rule for runners; deload on signals first, a planned lighter week acceptable).